Reinforcement learning (RL) is a critical stage in post-training large language models (LLMs), involving repeated interaction between rollout generation, reward evaluation, and centralized learning. Distributing rollout execution offers opportunities to leverage more cost-efficient inference resources, but introduces challenges in wide-area coordination and policy dissemination. We present ECHO-2, a distributed RL framework for post-training with remote inference workers and non-negligible dissemination latency. ECHO-2 combines centralized learning with distributed rollouts and treats bounded policy staleness as a user-controlled parameter, enabling […]

Ver mais

The study by Jung et al. (Jung H, Covino R, Arjun A, et al., Nat Comput Sci. 3:334-345 (2023)) introduced Artificial Intelligence for Molecular Mechanism Discovery (AIMMD), a novel sampling algorithm that integrates machine learning to enhance the efficiency of transition path sampling (TPS). By enabling on-the-fly estimation of the committor probability and simultaneously deriving a human-interpretable reaction coordinate, AIMMD offers a robust framework for elucidating the mechanistic pathways of complex molecular processes. This commentary provides a discussion […]

Ver mais

Structural shifts pose a significant challenge for graph neural networks, as graph topology acts as a covariate that can vary across domains. Existing domain generalization methods rely on fixed structural augmentations or training on globally perturbed graphs, mechanisms that do not pinpoint which specific edges encode domain-invariant information. We argue that domain-invariant structural information is not rigidly tied to a single topology but resides in the consensus across multiple graph structures derived from topology and feature similarity. To […]

Ver mais

Motivated by the growing interest in representation learning approaches that uncover the latent structure of high-dimensional data, this work proposes new algorithms for reconstruction-based manifold learning within Reproducing-Kernel Hilbert Spaces (RKHS). Each observation is first reconstructed as a linear combination of the other samples in the RKHS, by optimizing a vector form of the Representer Theorem for their autorepresentation property. A separable operator-valued kernel extends the formulation to vector-valued data while retaining the simplicity of a single scalar […]

Ver mais

arXiv:2601.22231v1 Announce Type: new Abstract: This paper revisits the role of positional embeddings (PEs) within vision transformers (ViTs) from a geometric perspective. We show that PEs are not mere token indices but effectively function as geometric priors that shape the spatial structure of the representation. We introduce token-level diagnostics that measure how multi-view geometric consistency in ViT representation depends on consitent PEs. Through extensive experiments on 14 foundation ViT models, we reveal how PEs influence multi-view geometry and […]

Ver mais

Nighttime vehicle detection poses significant challenges due to reduced visibility, uneven illumination, and increased noise in low-light imagery. While deep learning approaches have achieved remarkable success in daytime scenarios, their application to nighttime conditions remains constrained by the scarcity of specialized datasets and the computational demands of existing architectures. This paper presents three primary contributions to address these challenges. First, we introduce the Low-light Vehicle Annotation Dataset (L-VAD), comprising 13,648 annotated frames captured exclusively during nighttime conditions across […]

Ver mais

Large Language Models (LLMs) often produce hallucinated or unverifiable content, undermining their reliability in factual domains. This work investigates Reinforcement Learning with Verifiable Rewards (RLVR) as a training paradigm that explicitly rewards abstention ("I don’t know") alongside correctness to promote intellectual humility. We fine-tune and evaluate Granite-3.3-2B-Instruct and Qwen-3-4B-Instruct on the MedMCQA and Hendrycks Math benchmarks using a ternary reward structure ($-1$, r_abs, 1) under varying abstention reward structures. We further study the effect of combining RLVR with […]

Ver mais

deepseek dropped a paper on manifold constrained hyper connections (mhc) on jan 1st. liang wenfeng is a coauthor. paper: https://www.arxiv.org/abs/2512.24880 the basic idea: as models scale, letting different parts share more information internally helps performance but causes instability. mhc constrains this sharing to preserve stability while still getting the benefits. counterpoint research called it a “striking breakthrough” for scaling. omdia analyst said it could have ripple effects across the industry. what interests me is the timing. theres been […]

Ver mais